From 34840679695b8f230b404cd524408794ed4baf9b Mon Sep 17 00:00:00 2001
From: Petr Baudis <pasky@ucw.cz>
Date: Thu, 11 Mar 2010 23:33:18 +0100
Subject: [PATCH] tex: Go Styles random updates

---
 tex/gostyle.tex | 31 +++++++++++++++++++++++++------
 1 file changed, 25 insertions(+), 6 deletions(-)

diff --git a/tex/gostyle.tex b/tex/gostyle.tex
index 31e41d4..dfce3b5 100644
--- a/tex/gostyle.tex
+++ b/tex/gostyle.tex
@@ -817,7 +817,7 @@ based on the similarity of their pattern vectors,
 as well as discover correlations between styles and proportions
 of played patterns.
 
-Experts were asked to mark each style aspect of the given players
+Experts were asked to mark four style aspects of each of the given players
 on the scale from 1 to 10. The style aspects are defined as shown:
 
 \vspace{4mm}
@@ -838,6 +838,15 @@ Thickness $\theta$ & Safe & Shinogi \\ \hline
 %\end{table}
 \vspace{4mm}
 
+We have devised these four style aspects based on our own Go experience
+and consultations with other experts.
+The used terminology has quite
+clear meaning to any experienced Go player and there is not too much
+room for confusion, except possibly in the case of ``thickness'' ---
+but the concept is not easy to pin-point succintly and we also did not
+add extra comments on the style aspects to the questionnaire deliberately
+to accurately reflect any diversity in understanding of the terms.
+
 Averaging this expert based evaluation yields \emph{reference style vector}
 $\vec s_r$ (of dimension $4$) for each player $r$
 from the set of \emph{reference players} $R$.
@@ -973,6 +982,7 @@ Eigenval. & $\tau$ & $\omega$ & $\alpha$ & $\theta$ & Year \\
 \begin{table}[!t]
 % increase table row spacing, adjust to taste
 \renewcommand{\arraystretch}{1.3}
+\begin{threeparttable}
 \caption{Characteristic Patterns of PCA Dimensions}
 \label{fig:style_patterns}
 \centering
@@ -980,7 +990,7 @@ Eigenval. & $\tau$ & $\omega$ & $\alpha$ & $\theta$ & Year \\
 % than the plain LaTeX2e tabular which is used here.
 \begin{tabular}{|cccc|}
 \hline
-PCA1 top &
+PCA1 correl. \tnote{1} &
 \begin{psgopartialboard*}{(8,1)(12,6)}
 \stone[\marktr]{black}{k}{4}
 \end{psgopartialboard*} &
@@ -993,8 +1003,8 @@ PCA1 top &
 \stone[\marktr]{black}{j}{4}
 \end{psgopartialboard*} \\
 $0.447 \cdot$ & $0.274$ & $0.086$ & $0.083$ \\
-& side extension or \par 4--4 corner opening & high corner approach & high distant pincer \\
-PCA1 bot. &
+& high corner/side opening & high corner approach & high distant pincer \\
+PCA1 anticor. \tnote{2} &
 \begin{psgopartialboard*}{(3,1)(7,6)}
 \stone{white}{d}{4}
 \stone[\marktr]{black}{f}{3}
@@ -1012,12 +1022,18 @@ $0.447 \cdot$ & $-0.399$ & $-0.399$ & $-0.177$ \\
 & low corner approach & low corner reply & low corner enclosure \\
 \hline
 \end{tabular}
+\begin{tablenotes}
+\item [1] The more frequent the shape, the more moyo-oriented and thin-playing the player.
+\item [2] The more frequent the shape, the more territorial and thick-playing the player.
+\end{tablenotes}
+\end{threeparttable}
 \end{table}
 
 It is immediately
 obvious both from the measured $r$ and visual observation
 that by far the most significant vector corresponds very well
-to the territoriality of the players,\footnote{Cho Chikun, perhaps the best-known
+to the territoriality of the players,%
+\footnote{Cho Chikun, perhaps the best-known
 territorial player, is not well visible in the cluster, but he is
 positioned around $-0.8$ on the first dimension.}
 confirming the intuitive notion that this aspect of style
@@ -1063,7 +1079,8 @@ of neural network (sec. \ref{neural-net}) and $k$-NN classifiers (sec. \ref{knn}
 
 To compare and evaluate both methods, we have performed $5$-fold cross validation
 and compared their performance with a~random classificator.
-In the $5$-fold cross-validation, we randomly divide the training set into $5$ distinct parts with comparable
+In the $5$-fold cross-validation, we randomly divide the training set
+(organized by players) into $5$ distinct parts with comparable
 sizes and then iteratively use each part as a~testing set (yielding square error value), while
 the rest (remaining $4$ parts) is taken as a~training set. The square errors across all $5$ iterations are
 averaged, yielding mean square error.
@@ -1361,6 +1378,8 @@ Received BSc degree in Informatics at Charles University, Prague in 2009,
 currently a graduate student.
 Doing research in the fields of Computer Go, Monte Carlo Methods
 and Version Control Systems.
+Plays Go with the rank of 2-kyu on European tournaments
+and 2-dan on the KGS Go Server.
 \end{IEEEbiographynophoto}
 
 \begin{IEEEbiographynophoto}{Josef Moud\v{r}\'{i}k}
-- 
2.11.4.GIT